Custom Query (36 matches)

1. add f4 tests with invalid genotypes (to test problematic parsing, valgrind)
2. add f1 tests to see if introducing GeneProps? did not change the calculation of properties (and detect the 'm' bug!)
3. implement approximate, fast, very low and constant complexity coloring in f4 (styles) just like it is done in f1
4. oper_f4.cpp:137 maybe move all modifier definitions to one location
5. oper_f4.cpp:178 address this TODO
6. f4_general.cpp:478 address this TODO
7. "default move control to neuclasshandler" unclear meaning
8. same: "error: if no previous"
9. same: "Need to check if exists"
10. same: //they belong to neurons
11. same: "tc3 is only to ensure that in the end of neuron parameter definition"
12. check: sprintf((char*)buf2.c_str(), "%d", i1); and all other similar use cases
13. makeNull() never used?
14. char count; repetition counter - TODO discuss the CHAR type
15. f4_general.h from line 160 upwards. Fix English, merge (remove spaces) * (pointer) with variable names or type names when there is no variable name
16. unify modifier macros between f1 and f4?
17. still true? #define F4_MODIFIERS_VISUAL "" //not supported in f4
18. fix Geno_f4::style, old pieces of code remain and clash with new
19. still needed?

    	double       state;            ///<state of neuron
    	double       inertia;          ///<inertia of neuron
    	double       force;            ///<force of neuron
    	double       sigmo;            ///<sigmoid of neuron
    

20. out of line comment?

// a collection of cells, like Organism, for developmental encoding and another // convert f4 geno string to tree structure (internal)

21. unclear in parentheses:

    @return 0 if repair can be performed, -1 otherwise (the repair flag wasn't set in constructor)
    

22. int min, int max -> int mn, int mx
23. @param current parent of the analysed branch of the genotype - there is no "current" arg?
24. notation everywhere:

    double * v -> double *v
    const char * parse() -> const char* parse()
    

Compile new and old f4 sources, test all sample f4 genotypes from docs and see how their corresponding f0 differs (it should not).

Also test sequences of mutations and crossovers for days using valgrind or another similar tool(s). Introduce unit tests if needed - where you feel the source is tricky or very complex.

In particular, look at the following functions and structures and see how much code overlaps:

f4_Props::executeModifier(char modif)

f4_Props::f4_Props()

f4_Props::normalizeBiol4()

F1Props stdprops

Builder::grow(int part1, const char*g, Pt3D k, F1Props c)

Discuss what is different and whether it would be possible to unify these sources without losing backward compatibility in f4.